Current location: AGI. ASI not yet unlocked.
Most people are waiting for a “JARVIS moment” before they’ll believe AGI has arrived.
You know the scene. A calm voice in the ceiling that greets you by name, runs the house, argues with you about your calendar and is very obviously someone.
But that’s a scene from a movie, not a milestone anyone is actually measuring.
Meanwhile, the real milestone came and went. There wasn’t even a news headline about it, and most people haven’t noticed that it happened at all.
”AGI” Is Already Here
Let’s get the definitions out of the way first.
The AI industry defines AGI as above-average human intelligence across most tasks. My own definition differs slightly, as does that of most AI researchers: AGI is an AI with above-average human intelligence plus sentience and self-awareness.
But in the interests of corporate profits, industry has decided that AGI has already arrived.
Non-technical people overestimate AI because they picture JARVIS. Technical people underestimate it because they only see the limitations.
The truth is actually: broad capability, no autonomy yet.
And the numbers back this up.
On diagnostic reasoning, OpenAI’s o1 reached the correct diagnosis in 78% of studied cases. Physicians managed around 30%.
On competitive coding, only about 200 humans still out-rank the best models on Codeforces.
On forecasting, a Claude-based system placed 33rd out of 1,130 human forecasters in a Metaculus cup. That’s the top 3%.
Medicine, coding and forecasting. Three unrelated fields, and in each of them AI is now better than most of the humans who do it for a living.
I’ve seen this first-hand too.
A little while ago I set up AI coding agents to run unsupervised against a product’s entire backlog. Not one ticket at a time with me watching over its shoulder. The whole queue.
It cleared the entire backlog on its own in a day. That’s work that would have taken a developer days, or more likely, weeks.
And the newest generation of models, Claude Fable 5 and GPT Astra 6, are designed to do exactly this out of the box.
ASI Is Next, But Not the Movie Version
So if “AGI” is already here, what comes next?
ASI, or artificial superintelligence. The clean definition: AI that surpasses human ability in every field and can improve its own design without human input.
“AGI” is smarter than you and I. ASI is smarter then the world’s true geniuses and is capable of getting smarter on its own.
But we’re not there yet.
There’s a benchmark called Humanity’s Last Exam: 2,500 expert-level questions built specifically to exclude anything current AIs can already solve. The top model scores 48.4%. Human experts score around 90% in their own fields.
So no, we have not given birth to Skynet yet.
My favourite way to think about where we actually are is still droids from Star Wars.
Today’s AI is not C-3PO or R2-D2. It’s not a living, walking, talking robot with a personality that decides what to do next.
It’s more like an immobile droid head bolted to your desk that you have to prompt every single time.
Enormous capability, but no autonomy and no sentience.
What This Means for You
Media and industry reporting on all of this remains fragmented, so don’t try to keep up with it every day.
The reality is very simple: for most tasks, AI is now better than the average human.
That may not be “true AGI”, but it is worth setting aside a moment to think about.
What does that mean for the workforce? What does that mean for you?
My take — entry-level work is going to be further obliterated.
Business owners and CEOs are learning that it’s cheaper to pay API costs or a $200 Claude plan than to hire and train a graduate who knows nothing.
This is short-sighted, because in ten years’ time there will be nobody with the experience to step into the senior roles… but that is the current state of things.
And for those of us already in the workforce?
Your experience, plus the ability to use AI to force-multiply that experience, is still your biggest leverage. It’s what makes you valuable to your current employer and any future ones.
Two things to do this month.
First, pick one job you assumed needed a human expert, like a contract review, a financial model, or working out why your sales funnel is leaking, and test whether your current AI can already do it.
Second, write down the three places where your experience plus AI is your edge. Then go deeper there.
Courses
I have a range of online courses that teach business people what they need to know about productivity and AI:
1. The AI Playbook
This is my longer premium course on how businesses can deploy AI tools and technology across their processes and teams.
2. The Complete Claude, Claude Code & AI for Work Productivity
A comprehensive course taking you from beginner AI concepts like prompt and context engineering, to cutting-edge AI productivity using terminal-based AI tools like Claude Code for non-coding office work. Usable with ChatGPT, Claude, Gemini and other current LLMs.
3. The AI Agents, Automations and Agentic Workflows Guide
This non-technical course shows business people and non-coders how they can build and use AI agents in ChatGPT and Zapier.
4. ChatGPT for Managers
See how AI can solve complex management challenges in less than 30 seconds. Full prompt library and examples included.
5. Next Level Productivity
A practical, straightforward course that teaches you how to achieve elite-level personal productivity in today’s constantly interrupted world.
That’s it for this week!
— Aaron